Condition monitoring

A condition monitoring system, usually abbreviated CMS, watches the rotating machinery. Classically that means accelerometers on the main bearing, gearbox stages, and generator, sampled fast enough to resolve the characteristic frequencies of each component, plus oil particle counters and temperature sensors. Frequency-domain analysis then attributes energy at a given frequency to a specific bearing race or gear mesh.

Dedicated hardware of this kind is precise and expensive, and it is not present on every machine. Much of an operating fleet, particularly older onshore assets, carries only the temperature and low-rate vibration channels already in SCADA data. The practical question for most operators is therefore how much can be established from what the fleet already records, and the answer is more than is generally assumed.

Ten-minute temperature and power data will not resolve a bearing frequency, but it will reveal that a component now runs warmer than it used to for the same load and ambient conditions. That relationship drifting is the earliest broadly available signal of gearbox failure and of generator and main bearing degradation, and it is available on every turbine without fitting anything.

Condition monitoring answers a different question from performance analysis and the two are complementary. Performance work asks whether a machine is producing what it should; condition work asks whether it will still be running next quarter. A fleet needs both, and a finding that carries both readings - energy at stake and risk of failure - is what makes predictive maintenance plannable.